T032: Compound activity: Proteochemometrics

.github/workflows/ci.yml

@@ -76,11 +76,12 @@ jobs:
         shell: bash -l {0}
         run: |
           PYTEST_ARGS="--nbval-lax --current-env --dist loadscope --numprocesses 2"
-          if [ "$RUNNER_OS" != "Windows" ]; then
-            pytest $PYTEST_ARGS teachopencadd/talktorials/T*/talktorial.ipynb
-          else
-            pytest $PYTEST_ARGS teachopencadd/talktorials/ --ignore=teachopencadd/talktorials/T008_md_simulation/talktorial.ipynb --ignore=teachopencadd/talktorials/T019_md_simulation/talktorial.ipynb
-          fi
+          pytest $PYTEST_ARGS teachopencadd/talktorials/T032_compound_activity_proteochemometrics/talktorial.ipynb
+          # if [ "$RUNNER_OS" != "Windows" ]; then
+          #   pytest $PYTEST_ARGS teachopencadd/talktorials/T*/talktorial.ipynb
+          # else
+          #   pytest $PYTEST_ARGS teachopencadd/talktorials/ --ignore=teachopencadd/talktorials/T008_md_simulation/talktorial.ipynb --ignore=teachopencadd/talktorials/T019_md_simulation/talktorial.ipynb
+          # fi
 
   format:
     name: Black

devtools/test_env.yml

@@ -13,37 +13,39 @@ dependencies:
   # Explicitly add numpy because of https://github.com/volkamerlab/teachopencadd/issues/150
   - numpy
   - scikit-learn
+  - scipy
   # API changed after v2.6, see https://github.com/volkamerlab/teachopencadd/issues/265
-  - tensorflow<=2.6
+  #- tensorflow<=2.6
   - seaborn
-  - matplotlib-venn
+  #- matplotlib-venn
   # Remove jsonschema once this issue is fixed: https://github.com/Yelp/bravado/issues/478
   - jsonschema<4.0.0
-  - bravado
-  - requests
-  - requests-cache
+  #- bravado
+  #- requests
+  #- requests-cache
   - redo
-  - suds-community
-  - beautifulsoup4
-  - chembl_webresource_client
-  - pypdb
+  #- suds-community
+  #- beautifulsoup4
+  #- chembl_webresource_client
+  #- pypdb
   - biopython<=1.77
-  - biopandas
+  #- biopandas
   - rdkit==2021.09.5
-  - openbabel
-  - opencadd
-  - biotite>=0.34.0
-  - smina
-  - mdanalysis>=1.0.0
-  - mdtraj
-  - plip
-  - openmm
+  #- openbabel
+  #- opencadd
+  #- biotite>=0.34.0
+  #- smina
+  #- mdanalysis>=1.0.0
+  #- mdtraj
+  #- plip
+  #- openmm
   # depends on openff-toolkit->ambertools -> not available on Windows yet!
   # - openmmforcefields
-  - pdbfixer
-  - tqdm
-  - lxml
-  - kissim
+  #- pdbfixer
+  #- tqdm
+  #- lxml
+  #- kissim
+  - mordred
   ## CI tests
   # Workaround for https://github.com/computationalmodelling/nbval/issues/153
   - pytest 5.*
@@ -68,5 +70,11 @@ dependencies:
       - black-nb
       - nbsphinx-link
       - sphinxext-opengraph
+      # T032 - we might move some of this to conda-forge
+      - https://github.com/OlivierBeq/Papyrus-scripts/tarball/master
+      - prodec
+      - rich-msa
+      # Dependency for ClustalO webservice (also conda installable via -c bioconda)
+      - xmltramp2
       # TeachOpenCADD itself
       - ../

docs/all_talktorials.rst

@@ -34,3 +34,4 @@ This is the complete list of talktorials available for online reading. Take into
    talktorials/T026_kinase_similarity_ifp.nblink
    talktorials/T027_kinase_similarity_ligand_profile.nblink
    talktorials/T028_kinase_similarity_compare_perspectives.nblink
+   talktorials/T032_compound_activity_proteochemometrics.nblink

docs/talktorials.rst

@@ -63,6 +63,7 @@ The basis for computer-aided drug discovery
    talktorials/T013_query_pubchem.nblink
    talktorials/T021_one_hot_encoding.nblink
    talktorials/T022_ligand_based_screening_neural_network.nblink
+   talktorials/T032_compound_activity_proteochemometrics.nblink
 
 Structural biology
 ------------------

teachopencadd/talktorials/T032_compound_activity_proteochemometrics/README.md

@@ -0,0 +1,57 @@
+# T032 · Compound activity: Proteochemometrics
+
+**Note:** This talktorial is a part of TeachOpenCADD, a platform that aims to teach domain-specific skills and to provide pipeline templates as starting points for research projects.
+
+Authors:
+
+- Marina Gorostiola González, 2022, Computational Drug Discovery, Drug Discovery & Safety Leiden University (The Netherlands)
+- Olivier J.M. Béquignon, 2022, Computational Drug Discovery, Drug Discovery & Safety Leiden University (The Netherlands)
+- Willem Jespers, 2022, Computational Drug Discovery, Drug Discovery & Safety Leiden University (The Netherlands)
+
+
+
+## Aim of this talktorial
+
+While activity data is very abundant for some protein targets, there are still a number of underexplored proteins where the use of machine learning (ML) for activity prediction is very difficult due to the lack of data. This issue can be solved leveraging similarities and differences between proteins. In this talktorial, we use Proteochemometrics modeling (PCM) to enrich our activity models with protein data to predict the activity of novel compounds against the four adenosine receptor isoforms (A1, A2A, A2B, A3).
+
+
+### Contents in *Theory*
+
+* Proteochemometrics (PCM) modeling
+* Data preparation
+    * Papyrus dataset
+    * Molecule encoding: molecular descriptors
+    * Protein encoding: protein descriptors
+* Machine learning principles: regression
+    * Data splitting methods
+    * Regression evaluation metrics
+    * ML algorithm: Random Forest
+* Applications of PCM in drug discovery
+
+
+### Contents in *Practical*
+
+* Download Papyrus dataset
+* Data preparation
+    * Filter activity data for targets of interest
+    * Align target sequences
+    * Calculate protein descriptors
+    * Calculate compound descriptors
+* Proteochemometrics modeling
+    * Helper functions
+        * Preprocessing
+    * Model training and validation
+        * Random split PCM model
+        * Random split QSAR models
+        * Leave one target out split PCM model
+
+
+### References
+
+* Papyrus scripts [GitHub](https://github.com/OlivierBeq/Papyrus-scripts)
+* Papyrus dataset preprint: [<i>ChemRvix</i> (2021)](https://chemrxiv.org/engage/chemrxiv/article-details/617aa2467a002162403d71f0)
+* Molecular descriptors (Modred): [<i>J. Cheminf.</i>, 10, (2018)](https://jcheminf.biomedcentral.com/articles/10.1186/s13321-018-0258-y)
+* Protein descriptors (ProDEC) [GitHub](https://github.com/OlivierBeq/ProDEC)
+* Regression metrics [(Scikit learn)](https://scikit-learn.org/stable/modules/model_evaluation.html#regression-metrics)
+* XGBoost [Documentation](https://xgboost.readthedocs.io/en/stable/index.html)
+* Proteochemometrics review: [<i>Drug Discov.</i> (2019), <b>32</b>, 89-98](https://www.sciencedirect.com/science/article/pii/S1740674920300111?via%3Dihub)

teachopencadd/talktorials/T032_compound_activity_proteochemometrics/T032_env.yml

@@ -0,0 +1,27 @@
+name: teachopencadd_t032
+channels:
+  - conda-forge
+  - defaults
+dependencies:
+  - python>=3.8
+  - pip
+  - jupyter
+  - jupyterlab>=3
+  - nglview>=3
+  - pandas
+  - numpy
+  - biopython<=1.77
+  - rdkit==2021.09.5
+  - scikit-learn
+  - scipy
+  - seaborn
+  # Dependencies for PCM and papyrus scripts
+  - mordred
+  - pytest
+  - pip:
+      - https://github.com/OlivierBeq/Papyrus-scripts/tarball/master
+      - prodec
+      - rich-msa
+      # Dependency for ClustalO webservice (also conda installable via -c bioconda)
+      - xmltramp2
+

teachopencadd/talktorials/T032_compound_activity_proteochemometrics/data/README.md

@@ -0,0 +1,9 @@
+# T032 · Compound activity: Proteochemometrics
+## Data
+
+This folder stores input and output data for the Jupyter notebook.
+
+- `papyrus`: Directory with Papyrus bioactivity dataset downloads.
+- `sequences.fasta`: Sequences of the targets of interest for PCM modelling, in FASTA format.
+- `aligned_sequences.aln-fasta.fasta`: ClustalO multiple sequence alignment output, in FASTA format.
+- `aligned_sequences.[...]`: Additional ClustalO output files, not needed for the talktorial. 

teachopencadd/talktorials/T032_compound_activity_proteochemometrics/images/README.md

@@ -0,0 +1,8 @@
+# T032 · Compound activity: Proteochemometrics
+
+## Images
+
+This folder stores images used in the Jupyter notebook.
+- `PCM_model_text-01.png`
+- `papyrus_workflow.png`
+- `splitting_methods.png`

teachopencadd/talktorials/T032_compound_activity_proteochemometrics/scripts/README.md

@@ -0,0 +1,6 @@
+# T032 · Compound activity: Proteochemometrics
+## Scripts
+
+This folder stores scripts needed for the Jupyter notebook.
+
+- `clustalo.py`: ClustalO REST API client.